The crudest way is to eyeball them, if not, think about what defines a trend for you, and look for markets that fit those criteria more than others, it's not that complicated, you just have to decide what you're after!
The thing is you just can't make a difference between a simple volatility effect and the real underlying trend. Just to convince you, simulate a random walk with a fixed trend and a fixed volatility. You'll see a lot of very decent small trends in your signal, but you know perfectly (as you are the 'god' who simulated the signal) that they are not real trends. This doesn't cancel the fact you can try to make money out of volatility if there are some short term autocorrelations.
the existence of a "trend" I think would depend on the timescale you are interested in: hours, days, weeks, months, years. correct me if I'm wrong but I think the difference between chop and trend is whether there are dominant spectral components whose wavelengths are comparable to or longer than your timescale of interest. I suppose one way to find out is to run an FFT on a suitable duration of your timeseries of interest... and hope that the timeseries of the immediate past will reveal the correct spectrum of the future timeperiod during which you will be trading. Of course the spectrum will change over time so you will have to continually keep re-estimating...
I didn't think FFT's were the analytic method of choice as they don't meet the stationary constraints?